Well, satellite pics alone won’t do the trick.
The way we do it is to use a sat overlay which displays the temperature of the cloud top in the area you want to know. One image which works for this is the IR10.8 imaging at Flugwetter.de. I don’t know any public domain pics which display CLT, but I am sure there must be some. I’ll check if I can find something useful.
Then you get yourself a sounding from the same area and plot the temperature vs altitude. e.g. from here: http://weather.uwyo.edu/upperair/sounding.html
To see a sounding from this site, you need to define GIF:SKEW T for the plot, then the station on the map.
This works pretty well.
The GRAMET and other forecast products are of course exactly that, forecasts. So while they may be accurate, they are not nowcasting.
Thanks a lot. Will try that tonight! I do have a PCmet account!
As Mooney_Driver says the general idea is to get an infrared image and deduct the temperature of the highest cloud layer. Then from another source you need the vertical temperature profile to translate this temperature into an altitude. But there are limitations which means this will always be a rough guess only. Precise temperature profiles are not available, only sounding or predictions which will never be perfect. Then the temperatur of the upper cloud surface can be a few degrees of from the temperature of the sourrunding free air. Also if there is a temperatre inversion the temperature to altitude mapping may not be possible.
In fact the infrared layer in my ADLConnect iPad app does exactly that. It takes the IR image, transforms it into temperatures and then transforms that into altitudes based on the GFS temperature model. Unfortunately the satellite infrared images used require a flugwetter.de account so I can not make it available on a free website.
Sebastian,
I have the ADL, the app, and the account – but I did not yet discover how to read those infrared images. Could you show me an example?
This manual might help. It also contains some sample images:
Eumetsat provides a Cloud Top Height product.
Unfortunately it’s only available as full disk so it’s difficult to use against a route.
Here is a guide on how to use sat images for that. That fellow pilot 
likes to use the “visible” sat images, whilst pilots I know prefer the infrared version.
That’s my site, but I have never used visible images for cloud tops. Only IR. The URL is “IR-SAT” after all.
I had a look at the products available and will run this by our web guys at some stage.
The actual IR products which are available publicly (even the one in flugwetter) are of a rather large scale and may therefore be difficult to use for this.
At work, with the Ninjo Workstations, we do actually have a mouseover function, which allows to simply move the mouse over the IR picture and directly read the cloud temperature off the display. That helps enormeously when trying to do that quick. We also have much better sounding representation than the Wyoming temps, maybe someone might take their raw data once and make a nicer one available, also with the function that you could mouse over the thing and get a data readout. The one in Flugwetter is better.
One bit I keep hearing about the soundings are that they are outdated immediately when published, as the values are about an hour old. Yes and no.
Yes, by the time a sounding reaches you, it will be x hours old. In most weather situations, upper air temperatures will however stay pretty stable. If you however are looking at a frontal movement or so, you can use a sounding of a station where the front is now and compare the temperature profile with the local one. Cloud tops in many fronts stay in a certain ballpark as the front moves, unless it hits mountain ranges.
I have been lobbying for a really good cloud top height forecast product for years. Actually, the only one I am aware of is the TEMSI/EUROC which is very low resolution, yet I hope the new Alpfor products which are in development may get some of this. With the high res models of today, cloud top height should be much easier to predict.
Peter,
yeah sorry, should have said “monochromatic” vs. “RGB”.
